美国为削弱在 AI 关键投入品上对中国的依赖，正推进新的多国供应链联盟，重点锁定芯片与关键矿物。12 月 12 日，美国将与日本、韩国、新加坡、荷兰、英国、以色列、阿联酋与澳大利亚举行峰会，目标在能源、关键矿物、先进制造、半导体、AI 基础设施与运输物流上达成协议。此举延续自首次特朗普政府以来的供应链重组路径，但强调 AI 产业“全层级”布局，而非仅限矿物单一环节。中国在稀土链中的主导地位依旧压倒性：其炼制能力占全球逾 90%，排名第二的马来西亚仅 4%，令美国长期难以摆脱结构性依赖。

新计划的核心是聚焦“生产国”而非广泛结盟，并由 36 岁的美国国务院经济事务次长 Jacob Helberg 主导，他强调 AI 竞争已是“美中双马赛”，各国加入原因在于 AI 对经济规模与军事能力的影响。此前，美国曾建立能源资源治理倡议（特朗普时代）与矿产安全伙伴关系（拜登时代），但均未能撼动中国在稀土精炼上的垄断。近期中国虽在 10 月宣布加强稀土出口管控，但在特朗普与习近平会晤后同意暂停一年。

Helberg 强调此次合作是“美国中心”战略，旨在确保美企能不受“胁迫性依赖”影响而持续构建变革性技术。相关国家因掌握领先的半导体企业、关键矿产资源或战略产业地位而被选入，显示美国正以更紧凑、更产业链导向的框架巩固其在 AI 时代的供应链安全布局。

The US is advancing a new multinational supply-chain strategy to reduce dependence on China for AI-critical inputs, especially chips and minerals. On Dec. 12, the US will meet with Japan, South Korea, Singapore, the Netherlands, the UK, Israel, the UAE, and Australia to seek agreements spanning energy, critical minerals, advanced manufacturing, semiconductors, AI infrastructure, and logistics. The effort extends earlier Trump-era mineral-security programs but shifts from single-layer mineral focus to “full-stack AI” supply-chain architecture. China still dominates rare-earth processing with over 90% of global refining capacity versus Malaysia’s 4%, leaving US and allied efforts unable to dislodge entrenched dependence.

The new plan centers on “producer countries,” led by 36-year-old Undersecretary Jacob Helberg, who frames AI competition as a “two-horse race” between the US and China. Prior initiatives — the Trump Energy Resource Governance Initiative and Biden’s Minerals Security Partnership — improved coordination but did not overcome China’s structural lead. China briefly tightened rare-earth export controls in October before pausing them for one year following a Trump–Xi meeting.

Helberg describes the initiative as an “America-centric” strategy aimed at ensuring US firms can grow transformative technologies without exposure to coercive dependencies. Participating allies were chosen for semiconductor leadership, mineral resources, or strategic industrial roles, indicating a more focused and production-driven approach to securing US AI-era supply chains.